ORACLE CONFIDENTIAL. For authorized use only. Do not distribute to third parties.
Pre-General Availability: 2025-12-16
Manage LLM
Agent Factory supports a diverse selection of Large Language Models (LLMs) from multiple providers. Use the LLM Management screen to set up and manage both generative and embedding models within the application.
Generative Models
Agent Factory supports generative models from OCI Generative AI, Ollama, vLLM, and OpenAI. Use these models for tasks like text generation, summarization, question answering, and supporting MCP server tools.
OCI Gen AI
Agent Factory accepts many generative models available through the OCI Gen AI service. Commonly used models include:
cohere.command-r-plus-08-2024cohere.command-a-03-2025xai.grok-3xai.grok-4
Oracle recommends using xai.grok-4 model.
OpenAI
The following models are currently supported:
gpt-4ogpt-4o-mini
Ollama/vLLM
You can connect to any self-hosted model endpoint. Oracle recommends using Llama 3.3.
Use any of the following Wayflow-hosted models in your vLLM model configurations.
| Model | URL | MODEL_ID | Capabilities | Speed | Accuracy |
|---|---|---|---|---|---|
| Llama 3.1 8B | llama.wayflow.oraclecorp.com |
meta-llama/Meta-Llama-3.1-8B-Instruct |
Text only | Medium-High | Low |
| Llama 3.3 70B | llama70bv33.wayflow.oraclecorp.com |
/storage/models/Llama-3.3-70B-Instruct |
Text only | Medium | Medium-Low |
| Llama- 4 Maverick | llama4maverick.wayflow.oraclecorp.com |
Llama-4-Maverick |
Text only | Medium | Medium |
| Gemma-3-27b-it | gemma27b-it-v3.wayflow.oraclecorp.com |
google/gemma-3-27b-it |
Text and image inputs | Medium | Medium |
Note: Provided vLLM models work out of the box-no setup needed.
Embedding Models
Agent Factory supports the following embedding models. Use these models for transforming text into numerical vectors, enabling semantic search, and retrieval-augmented generation (RAG).
Local Models
The following pre-trained sentence transformer models are bundled with the application and run locally.
multilingual-e5-base(768 dimensions)all-MiniLM-L6-v2
OCI Gen AI
The following Cohere embedding models from OCI Gen AI are supported:
cohere.embed-english-light-v3.0(384 dimensions)cohere.embed-multilingual-light-v3.0(384 dimensions)cohere.embed-english-v3.0(1024 dimensions)cohere.embed-multilingual-v3.0(1024 dimensions)
Ollama/vLLM
You can use any embedding model hosted on an Ollama or vLLM endpoint as needed.
Configure an OCI Gen AI Model
-
Select Add Configuration button
-
In the Add New Configuration form, select Generative Model
-
Select OCI GenAI
-
Enter Configuration Name
-
Use the following details (for Oracle internal use) as a template for your configuration:
-
Model ID:
cohere.command-r-plus-08-2024 -
Endpoint:
https://inference.generativeai.us-chicago-1.oci.oraclecloud.com -
Compartment ID:
ocid1.compartment.oc1..aaaaaaaa..............vdyx7l3rn2gkfjgltdfw347fw5bq -
User:
ocid1.user.oc1..aaaaaaaap..............azw6gpinaukrb5p22geiuivjgpta -
Fingerprint:
8f:06:3d:b1:..............:fe:e0:44:c1:dd:9c:16 -
Tenancy:
ocid1.tenancy.oc1..aaaaaaaaw..............ha6uqms2h2ovxhcgwgmbl3dukqsjxa -
Region:
us-chicago-1 -
Key File: Upload your private API key file
-
-
Select Test Connection
-
Once the connection is successful, select Save Configuration
Configure an Ollama Model
-
Click Add Configuration button
-
In the Add New Configuration form, select Generative Model
-
Click Ollama tab
-
Enter Configuration Name
-
Below is the Ollama Model configuration that can be used within Oracle:
-
Model ID:
llama3.2 -
URL:
phoenix95898.dev3sub2phx.databasede3phx.oraclevcn.com -
PORT:
11433
-
-
Click Test Connection button
-
Once the connection is successful, click Save Configuration button